SUMO Protease SMT7 Modulates Ribosomal Protein L30 and Regulates Cell-Size Checkpoint Function

Proliferating cells actively coordinate growth and cell division to ensure cell-size homeostasis; however, the underlying mechanism through which size is controlled is poorly understood. Defect in a SUMO protease protein, suppressor of mat3 7 (SMT7), has been shown to reduce cell division number and increase cell size of the small-size mutant mating type locus 3-4 ( mat3-4 ), which contains a defective retinoblastoma tumor suppressor-related protein of Chlamydomonas ( Chlamydomonas reinhardtii ). Here we describe development of an in vitro SUMOylation system using Chlamydomonas components and use it to provide evidence that SMT7 is a bona ﬁ de SUMO protease. We further demonstrate that the SUMO protease activity is required for supernumerous mitotic divisions of the mat3-4 cells. In addition, we identi ﬁ ed RIBOSOMAL PROTEIN L30 (RPL30) as a prime SMT7 target and demonstrated that its SUMOylation is an important modulator of cell division in mat3-4 cells. Loss of SMT7 caused elevated SUMOylated RPL30 levels. Importantly, overexpression of the translational fusion version of RPL30-SUMO4, which mimics elevation of the SUMOylated RPL30 protein in mat3-4 , caused a decrease in mitotic division and recapitulated the size-increasing phenotype of the smt7-1 mat3-4 cells. In summary, our study reveals a novel mechanism through which a SUMO protease regulates cell division in the mat3-4 mutant of Chlamydomonas and provides yet another important example of the role that protein SUMOylation can play in regulating key cellular processes, including cell division.


INTRODUCTION
Cell size homeostasis in proliferating cells is an evolutionarily conserved trait Umen, 2005;Tzur et al., 2009;Goudarzi and Lindström, 2016). Cell size control requires coordination of growth and the cell cycle and until now, the underlying mechanism has only been extensively investigated in yeasts. Studies of yeasts have provided crucial evidence that the regulatory topology required for size control is similar to that found in the opisthokont branch of eukaryotes (Cross et al., 2011). In budding yeast, defects in Whiskey 5 (Whi5), the transcriptional inhibitor that controls G1/S transition, cause a small-cell phenotype (Jorgensen et al., 2002). A small-size phenotype is also observed in animals (opisthokonta branch) and the green alga Chlamydomonas (Chlamydomonas reinhardtii; Viridiplantae branch) cells with defective G1/S inhibitors, members of the retinoblastoma (Rb) family (Dannenberg et al., 2000;Sage et al., 2000;Fang et al., 2006) whose sequences are entirely different from Whi5. These findings suggest that the mode of action from similar regulatory circuits rather than homologous proteins have been conserved during evolution.
In addition to the G1/S transition network centered on Rb or Whi5, defects in ribosome biogenesis have been shown to affect cell size. In budding yeast (Saccharomyces cerevisiae), genomewide screening for size mutants revealed that defects in Split finger protein 1 (Spf1) transcription factor and Suppressor of cdc25 (Sch9) kinase that govern ribosome biogenesis and translation initiation generate small daughter cells Marion et al., 2004;Urban et al., 2007). However, the size threshold of yeasts is not static and is subject to changes in growth rate Ferrezuelo et al., 2012;Turner et al., 2012;Chica et al., 2016), a property that makes size control studies in yeasts complicated. It is extremely challenging to assess cell-size defects in multicellular organisms. Despite this, plant and animal cells within one tissue often display a remarkable uniformity in size (Lloyd, 2013;Ginzberg et al., 2015;Serrano-Mislata et al., 2015;Willis et al., 2016;Jones et al., 2017). Recent studies in animal cells reveal that cells adjust both cell cycle length and growth rate to maintain size homeostasis (Cadart et al., 2018;Ginzberg et al., 2018). Growth rate modulation controlled by ribosome-based protein translation has been suggested to regulate size homeostasis (Kafri et al., 2016). Even though deficiencies in the ribosome biogenesis pathway have been found to produce small cells in Drosophila (Drosophila melanogaster) and mouse (Mus musculus) systems (Montagne et al., 1999;Ruvinsky et al., 2005), evidence of a causal relationship between size homeostasis and ribosome-dependent regulation remains limited.
Protein conjugation by the small-ubiquitin-like modifier (SUMOylation) plays important roles during the cell division cycle (Li and Hochstrasser, 1999;Pelisch et al., 2014). Proteomic studies reveal that many cell cycle-related proteins are conjugated by SUMO Cubeñas-Potts et al., 2015). Indeed, SUMOylation has been found to be important in regulating several cell cycle proteins. For example, SUMO conjugation of cyclin E contributes to the control of replication firing (Bonne- Andrea et al., 2013). SUMO modification stabilizes cyclin dependent kinase 6 and regulates G1/S transition (Bellail et al., 2014). Forkhead box transcription factor M1 (FoxM1) is extensively SUMOylated and its SUMOylation is important for cell cycle progression and chromosome segregation (Schimmel et al., 2014). SUMO deconjugation enzymes human sentrinspecific protease 1 (SENP1) and SENP2 are important for precise spatial and temporal control of SUMOylation in mitosis (Cubeñas-Potts et al., 2013). Deregulation of these SUMOylation and deSUMOylation enzymes not only causes defects in cell proliferation and genome stability, but recent findings have also shown that different types of cancer cells require a functioning SUMOylation system (Eifler and Vertegaal, 2015).
When grown under light-dark cycles (e.g., 12 h light-12 h dark) mimicking the diurnal rhythm found in Nature, Chlamydomonas cells utilize modified cell cycle-multiple fission cell division. Multiple fission is characterized by a long G1 phase, during which cells can grow more than 10 times in size. At the end of G1, mother cells undergo successive rounds of alternating DNA synthesis (S phase) and mitosis (M phase) to produce 2 n daughters of uniform size. At early/mid G1, cells pass the first size checkpoint, "Commitment", which is equivalent to "Start" in budding yeast (Hartwell et al., 1974;Johnston et al., 1977) and "Restriction Point" in animals (Pardee, 1974;Dolznig et al., 2004), during which cells acquire sufficient mass to complete the cell cycle. The second size checkpoint occurs during S/M, where mother cells undergo a controlled number of division cycles to generate uniformly-sized daughters. The number of S/M cycles is dictated by the size of mother cells in which larger mother cells divide more times than smaller mother cells, so size homeostasis is achieved (Craigie and Cavaliersmith, 1982;Donnan and John, 1983). Multiple fission allows Chlamydomonas cell cultures to be naturally synchronized by an alternating light/dark regime. As a result, cells grow photosynthetically at G1 during the light period, while cell division cycle occurs in the dark. Because cellular functions associated with growth and cell division are temporarily separated, the daughter cell-size can be used as a direct readout to assess the size control mechanism (Umen, 2005;Fang et al., 2006).
The retinoblastoma (Rb) tumor suppressor pathway regulates Commitment size and cell division number in Chlamydomonas (Umen and Goodenough, 2001;Fang et al., 2006). Mutations in the Chlamydomonas Rb gene, MATING TYPE LOCUS 3 (MAT3), cause these cells to commit at a reduced cell size and to divide supernumerously at the S/M phase in a manner that generates small daughter cells (Umen and Goodenough, 2001). These defects are suppressed by mutations in E2F or DP, that encode subunits of evolutionarily conserved heterodimeric transcription factor E2F-DP, a regulatory duplex that acts directly downstream of the Rb pathway (Fang et al., 2006). In addition to E2F and DP, several extragenic partial suppressors of mat3 (SMTs) have been isolated (Fang and Umen, 2008;Fang et al., 2014). A defect in SMT7, which encodes a SUMO protease, in the mat3 mutant causes size suppression, rendering daughter cells (smt7-1 mat3-4) larger than mat3-4 but smaller than wild-type cells (Supplemental Figure 1). Interestingly, cells containing the single mutation, smt7-1, display no obvious size defect (Fang and Umen, 2008).
Here, we describe establishment of an in vitro Chlamydomonas SUMOylation system and provide biochemical evidence that SMT7 possesses SUMO protease activity. In addition, we show that SUMO protease activity of SMT7 is required for MAT3-dependent size checkpoint control. Ribosomal protein L30 (RPL30) was isolated as a SMT7-interacting protein whose SUMO deconjugation is regulated by SMT7. As a result of this interaction, a defect in SMT7 caused increased levels of RPL30 SUMOylation. Surprisingly, overexpression of RPL30-SUMO4 DGG -3XHA protein, which mimics SUMOylated RPL30 protein but not RPL30-3XHA protein in mat3-4 cells recapitulated smt7-1 mat3-4 cells and led to reduced cell division and size suppression. Together, our study provides unexpected insights into the size-mediated cell division cycle and demonstrates that SUMOylation of a ribosomal protein can have novel regulatory consequences.

Molecular Characterization of the SMT7 Locus
Even though a defect in a putative SUMO protease SMT7 has been demonstrated to suppress the small cell size of mat3-4 (Fang and Umen, 2008), the structure of SMT7 has not been fully characterized. Despite numerous attempts to amplify SMT7 cDNA, we failed to obtain the full-length cDNA. As an alternative, we combined RT-PCR and 39 rapid amplification of cDNA ends (RACE)-PCR to amplify overlapping SMT7 cDNA fragments (Supplemental Figure 2A) and validate the gene structure of SMT7 ( Figure 1A). SMT7 encodes a protein with a distinct N-terminal region followed by a conserved SUMO protease domain (Pfam 02902; Figure 1B). Protein sequence alignment of the SUMO protease domains of SMT7 and SUMO proteases from humans, Arabidopsis, and budding yeast indicated that the canonical catalytic triad (His 860 -Asp 877 -Cys 928 ) required for SUMO deconjugation function is evolutionarily conserved ( Figure 1C). Phylogenetic analysis revealed that SMT7 is related to SUMO proteases EARLY IN SHORT DAYS4 (ESD4) and its closest homologs (Supplemental Figure 2B). In addition to the SUMO protease domain, one potential nuclear localization sequence (NLS), and two putative SUMO-interacting motifs were identified in the SMT7 protein sequence (Supplemental Figure 3).

SUMO Protease Activity of SMT7 Is Required for Size Control in the mat3 Mutant
To detect and purify SMT7 protein, a C-terminal triple hemagglutinin (3XHA) epitope-tagged allele of a genomic SMT7 clone under the control of its native promoter and terminator was constructed (pSMT7-3XHA; Supplemental Figure 4). Immunoblotting with the anti-HA antibody confirmed that the expression of a major protein ran slightly larger than 100 kD, which is close to the predicted size (;106 kD) of full-length SMT7-3XHA protein in the smt7-1 mat3-4::SMT7-3XHA strains but not in the untagged wild-type strain ( Figure 2A). Expression of SMT7-3XHA protein was able to rescue the function of (A) Immunoblot detection of SMT7-3XHA protein with the anti-HA antibody from total protein extracts of three independent transgenic smt7-1 mat3-4 strains carrying the pSMT7.1-3XHA construct, mat3-4, smt7-1, and wild type (21gr) strains. SMT7-3XHA protein was shown as a single band predicted at the molecular mass ;106 kD. Immunoblot with Ribulose bisphosphate carboxylase large chain (RbcL) antibody was used as a loading control. (B) Immunoblot detection of SMT7 C928A -3XHA or SMT7 C928S -3XHA protein with the anti-HA antibody from total protein extracts of the independent transgenic smt7-1 mat3-4 strains carrying the pSMT7.1 C928A -3XHA or pSMT7.1 C928S -3XHA construct, respectively. Immunoblot with ATP synthase subunit beta (AtpB) antibody was used as a loading control. *A putative degraded SMT7-3XHA protein. Total protein extracts were separated by a 10% SDS-PAGE (A) or a 4% to 12% gradient gel (B). smt7-1 in the smt7-1 mat3-4 strain and restore the small-size phenotype of mat3 mutant in three independent transformants (Table 1).

SMT7 mRNA and Protein Are Diurnally Regulated
To investigate how SMT7 mRNA was regulated, RT-qPCR was initially used to monitor its expression in samples collected at twohour intervals from previously synchronous 21gr culture (Fang et al., 2014). However, the abundance of SMT7 mRNA was too low to be detected by quantitative PCR (qPCR). As an alternative, RTdroplet digital PCR (RT-ddPCR) was used to determine the expression pattern of SMT7 mRNA. RT-ddPCR has been shown to be more sensitive than RT-qPCR in detecting rare mRNAs (Hindson et al., 2013). SMT7 mRNA levels increased during the G1 phase, reached a peak at the G1/S transition, and then gradually declined as cells completed mitosis (Supplemental Figure 5). The expression pattern of SMT7 in the synchronized cell culture is similar to that previously reported by Zones et al. (2015).
Because our generated SMT7 antibody could not detect endogenous SMT7 protein, a smt7-1::SMT7-3XHA strain was generated by crossing the complemented smt7-1 mat3-4::SMT7-3XHA #3 to a wild-type (mt -, 6145C) strain. Cell size measurement of the segregated progeny confirmed again that expression of the SMT7-3XHA protein in the smt7-1 mat3-4 mutant restored the small cell size (Supplemental Table 1). The expression pattern of SMT7-3XHA protein in the smt7-1::SMT7-3XHA strain was determined every two hours by HA antibody in synchronized culture by immunoblotting. The culture synchrony was assessed by Commitment and mitotic index assays ( Figure 3A), and expression of S/M marker gene CDKB1 ( Figure 3B). To investigate dynamic changes of SMT7 protein either as the basis of overall cellular protein concentration or the amount of protein per cell, lysates were loaded on two blots, one with equal amounts of protein in each lane and one with equal numbers of cells in each lane ( Figure 3C). SMT7 protein was expressed throughout the cell cycle with its levels increased during G1, reaching the first peak at around the Commitment state, then declining before reaching the second peak at the S/M phase, which is marked by accumulation of the cell cycle marker CYCD3 protein (Bisová et al., 2005;Zones et al., 2015). SMT7 protein level was gradually decreased in postmitotic cells. The dual accumulation peaks were most prominent when equal numbers of cells were loaded for immunoblot analysis. Together, these results show that SMT7 mRNA and protein are regulated diurnally.
To gain further detailed information about where SMT7 accumulated in the nucleus, we used indirect immunofluorescence (IF) with anti-HA antibody. Even though the anti-HA antibody was able to successfully detect the SMT7-3XHA protein of smt7-1 mat3-4::SMT7-3XHA and smt7-1::SMT7-3XHA transgenic lines ( Figures  2 and 3) by immunoblotting, they failed to generate reliable IF signals in these complemented strains. To overcome this problem, a Glyrich linker was inserted between sequences encoding SMT7 and 3XHA tag to generate the pSMT7-lin-3XHA construct. Insertion of a flexible Gly-rich linker between a tagged protein and an epitope has been shown to increase epitope sensitivity and accessibility (Sabourin et al., 2007;Reddy Chichili et al., 2013). Expression of SMT7-lin-3XHA restored SMT7 function in the smt7-1 mat3-4 double mutant (Supplemental Table 2). Thus, pSMT7-lin-3XHA was introduced into the wild-type strain and the transgenic strains expressing the SMT7-lin-3XHA protein ( Figure 4B) were used for IF studies. Consistent with its expression in the nuclear fraction ( Figure 4A), SMT7-lin-3XHA signal was detected as a discontinuous rim-like staining of the nuclear periphery surrounding the 4',6-diamidino-2-phenylindole dihydrochloride (DAPI)-stained DNA typical of nuclear pore complex (NPC)-associated localization ( Figure 4C). The discontinuous rim-like staining of the nuclear periphery in SMT7-lin-3XHA was very similar to the distribution of NPC-associated SUMO protease in budding yeast and mammals (Zhang et al., 2002;Goeres et al., 2011;Chow et al., 2012). Hence, SMT7 is predominantly localized in nuclear envelopes and potentially targets to NPCs. Consistent but very weak cytoplasmic SMT7-lin-3XHA staining was also observed. Therefore, SMT7 may be present in the cytoplasm to some extent.

The SUMO Protease Domain of SMT7 Is Capable of Removing SUMO from SUMOylated Chlamydomonas Proliferating Cell Nuclear Antigen Protein In Vitro
To assess the SUMO protease activity of SMT7, an in vitro SUMOylation system was constructed. The Chlamydomonas SUMOylation components were identified using Arabidopsis SUMO activating enzyme 1 (AtSAE1), AtSAE2, E2 conjugating enzyme 1 (AtSCE1), AtSUMO1, and proliferating cell nuclear antigen (AtPCNA) proteins as the BLASTP queries for the Chlamydomonas proteome (see Methods) and the corresponding cDNAs were isolated by RT-PCR. The alignment of protein sequences of individual SUMOylation components are shown in Supplemental Figure 6. Among the SUMO genes we identified, SUMO4 mRNA is the most abundant SUMO gene transcript (Supplemental Figure 7A) and diurnally regulated (Zones et al., (A) Immunoblots of total cellular protein (whole cell) and the nuclear and cytoplasmic/membrane (Cyto/Mem) fractions prepared from smt7-1, and complemented smt7-1 strain expressing a triple HA-tagged SMT7 protein (smt7-1::SMT7-3XHA_I18). Histone 3 (H3) was used as a nuclear marker and CRD1 protein was used as a chloroplast membrane marker (Allen et al., 2008). Equal amounts of protein were loaded in each lane. The experiment was repeated three times, yielding an identical result, with representative data shown. (B) Immunoblots of total cellular protein from transgenic strains expressing triple HA-tagged SMT7 protein with linker (smt7-1::SMT7-lin-3XHA) in the wildtype (wt) and smt7-1 mat3-4 background. Immunoblots with RbcL antibody were used as a loading control. (C) Confocal immunofluorescence microscopy of wild-type and SMT7-lin-3XHA-expressing cells stained with anti-HA antibody and the DNA dye DAPI. High magnification image of the cell in red boxed region is shown. (D) Confocal immunofluorescence microscopy of smt7-1 mat3-4 and SMT7-lin-3XHA-expressing smt7-1 mat3-4 cells stained with anti-HA antibody and the DNA dye DAPI. High magnification image of the cell in red boxed region is shown. GFP/DAPI, merged signals from the GFP and DAPI channels; DIC, differential interference contrast images of cells. 2015). This gene was therefore chosen for the in vitro SUMOylation assay. The alignment of SUMO protein sequences from yeast, Chlamydomonas, Arabidopsis, and human are shown in Supplemental Figure 7B. Because proliferating cell nuclear antigen (PCNA) is known to be SUMO-conjugated in various organisms (Hoege et al., 2002;Leach and Michael, 2005;Gali et al., 2012;Moldovan et al., 2012;Strzalka et al., 2012) and at least one consensus SUMOylation site (K91) of Chlamydomonas PCNA can be predicted with high confidence by multiple prediction programs (Supplemental Figure 8), Chlamydomonas PCNA was used as the substrate for the SUMO conjugation test. Chlamydomonas SAE1 and SAE2 cDNAs were cloned into the pCDFDuet vector. Chlamydomonas SCE1 and SUMO4 were cloned into the pETDuet vector. Chlamydomonas PCNA was cloned into the pRSFDuet-3XHA vector. E. coli cells transformed with pCDFDuet-SAE1-SAE2, pETDuet-SCE1-SUMO4 GG , and pRSFDuet-3XHA-PCNA plasmids were selected ( Figure 5A; see Methods). Coexpression of SAE1, SAE2, SCE1, SUMO4, and 3XHA-PCNA was verified by immunoblotting ( Figure 5B). In addition to detecting a 3XHA-PCNA protein with the predicted size (;37 kD), a ;60 kD lowmobility SUMO4-conjugated 3XHA-PCNA protein was also detected when all of the SUMOylation components were present ( Figure 5B), indicating that PCNA protein is SUMOylated. Importantly, SUMO4 failed to be conjugated to PCNA when one or more of the SUMO conjugation components were missing. Interestingly, the E. coli endogenous proteins were also found to be conjugated with SUMO4 when all the SUMOylation components were present ( Figure 5B). Hence, we conclude that Chlamydomonas PCNA is SUMOylated and the in vitro SUMOylation system is established. (A) Expression vectors carrying the cDNAs encoding the individual SUMOylation components as indicated were generated and introduced into E. coli cells. (B) Chlamydomonas PCNA was conjugated with SUMO4. The SAE2 and SUMO4 GG are N-terminally fused to 6XHis. SAE1, SCE1, and PCNA are C-terminally fused to the S tag. PCNA is N-terminally fused to the 3XHA tag. (C) SUMOylated Chlamydomonas PCNA was deconjugated by SMT7 709-972 . Expression of each SUMOylation component was confirmed by immunoblotting using the corresponding antibodies. Expression of SMT7 was confirmed using SMT7-specific antibody. *SUMOylated PCNA. CBB staining showing equal loading of total protein amount. Immunoblotting showing two independent E. coli strains, depicted by "1" and "2", carrying the indicated constructs. Total protein extracts were separated on a 12% SDS-PAGE gel.
To test the SUMO protease activity of SMT7 and determine its ability to remove SUMO from SUMO-conjugated PCNA protein, SMT7 was co-expressed with the SUMOylation components along with 3XHA-PCNA protein. The full-length SMT7 cDNA is extremely GC-rich (;76%) and was challenging to clone. As an alternative, the SUMO protease domain of SMT7 (SMT7 709-972 ) was amplified and introduced into the pRSFDue-3XHA-PCNA plasmid to make pRSFDue-3XHA-PCNA-SMT7 709-972 . Expression of all the corresponding components was verified by immunoblotting ( Figure 5C). When SMT7 709-972 was expressed in the presence of all the SUMOylation components, only unconjugated 3XHA-PCNA was detected, indicating that SMT7 709-972 is capable of removing SUMO4 from conjugated PCNA. Coincidently, expression of SMT7 709-972 caused a decrease in levels of SUMO4 conjugates and an increase in levels of SUMO4 monomers in E. coli cell lysates ( Figure 5C). Expression of the defective SUMO protease domain, SMT7 709-972/C928A or SMT7 709-972/C928S , on the other hand, failed to remove SUMO4 from PCNA and SUMO4conjugated PCNA was detected ( Figure 5C). In conclusion, our data provide evidence that the SUMO protease domain of SMT7 possesses SUMO-deconjugating activity.

SUMO4-Conjugated Proteins Accumulate in the smt7-1 Mutant
To investigate whether SUMO4-conjugated proteins were regulated by SMT7 in vivo, SUMO4 antibody were raised to detect SUMO4-conjugates. The specificity of SUMO4 antibody was confirmed by immunoblotting of recombinant Chlamydomonas SUMO1, SUMO2, and SUMO4 proteins (Supplemental Figure 9) and the SUMO4 antibody was therefore used for immunoblotting Figure 6. Increased SUMO4 Conjugates in smt7-1 and smt7-1 mat3-4 Mutants.
of Chlamydomonas cellular extracts. In addition to SUMO4 monomer migrating close to ;13 kD, a larger number of SUMO4conjugated proteins were detected in the wild-type ( Figure 6A) and mat3-4 ( Figure 6B) strains. When SMT7 is missing, the levels of SUMO4 monomer were substantially decreased and the levels of the SUMO4-conjugated proteins were substantially increased, indicating that SMT7 is a major regulator of SUMO4 conjugates in Chlamydomonas.
Identification of Ribosomal Protein L30 as the Critical SMT7-Interacting Protein in mat3-4 To identify SMT7 targets and their roles in MAT3-mediated size checkpoint control, immunoprecipitation (IP) was conducted to isolate SMT7-interacting proteins. Because SMT7 and its targets may interact transiently, a substrate-trapping-based strategy was used to immunopurify SMT7-3XHA-interacting proteins from cellular extracts prepared from two independent smt7-1 mat3-4::SMT7 C928A -3XHA transgenic lines. Substrate-trapping takes advantage of the catalytic inactive mutant SMT7, SMT7 C928A -3XHA, which allows substrate binding but greatly slows its dissociation because of the impaired catalytic activity (Flint et al., 1997;Elmore et al., 2011), to enrich SMT7-interacting proteins. The smt7-1 mat3-4 mutant was used as a negative control for IP. IPs followed by liquid chromatography-mass spectrometry analysis identified SMT7 C928A -3XHA-interacting proteins as the following molecules: RPS2, RPS23, RPL30, a low-CO2-inducible protein, elongation factor 3a-related protein, light harvesting chlorophyll a/b binding protein 2 (Lhca2), a flagellar-associated protein, FtsH-like membrane ATPase/metalloprotease, plastid ribosomal protein S13, and a NAC domain containing protein (Table 2). Importantly, none of these proteins were detected in several parallel control IP experiments conducted using a smt7-1 mat3-4 strain lacking HA-tagged SMT7 C928A . Because RPL30 was isolated six out of ten times in independent IP experiments and ribosome biogenesis has been implicated in cell size control (Jorgensen et al., 2002;Gómez-Herreros et al., 2013), we chose to focus on RPL30.

SMT7 Interacts with RPL30 and Regulates the SUMOylation Status of RPL30
To confirm that SMT7 interacts with RPL30 in vivo, polyclonal antibody raised against RPL30 was generated. The purified RPL30 antibody recognized a ;12-kD band predicted to be the size of RPL30 protein in Chlamydomonas cell lysate ( Figure 7A). To validate that the antibody recognized endogenous RPL30, the anti-RPL30 antibody were used for IP from lysates prepared from the wild-type strain. Mass spectrometry analysis confirmed that RPL30 was pulled down in the IP pellets (Supplemental Figure 10; Supplemental Table 3). Hence, the specificity of anti-RPL30 antibody was validated.
We used the established in vitro SUMOylation assay to determine whether Chlamydomonas RPL30 can be SUMOylated and whether SMT7 is able to remove the SUMO-conjugated RPL30. Multiple SUMOylation prediction programs failed to identify a common SUMO acceptor Lys with high confidence in RPL30 (Supplemental Figure 8). Even so, when RPL30 was coexpressed with all of the SUMOylation components, an ;20-kD S-tag 3XHA-RPL30 protein and two slowly migrated RPL30 signals (;40 and ;50 kD) were detected by anti-HA antibody ( Figure 7B), suggesting two and three SUMO4 were conjugated to RPL30 protein in vitro. As expected, SUMO4-conjugated RPL30 proteins disappeared when SMT7 709-972 was co-expressed ( Figure 7B), indicating that the SUMO protease activity of SMT7 709-972 was able to remove SUMO4 from SUMOylated RPL30.
To test whether SUMOylation of RPL30 was regulated by SMT7 in vivo, RPL30 protein was examined in strains lacking SMT7 by immunoblotting. The levels of unconjugated RPL30 protein were slightly but consistently reduced in smt7-1 and smt7-1 mat3-4 strains compared with the wild-type and mat3-4 strains, respectively (Supplemental Figure 11A). A slow-mobility RPL30 signal running at ;22 kD that corresponds to the mono-SUMO conjugated RPL30 was repeatedly detected (Supplemental  (A) Immunoblot detection of HA-tagged SMT7 C928A protein from anti-RPL30 immunoprecipitation of smt7-1 mat3-4 and the complemented smt7-1 mat3-4::SMT7 C928A -3XHA_A22 strains. No antibody (No Ab) was used as a negative control for IP. Figure 11A). However, the polyclonal RPL30 antibody recognized many non-specific proteins and worked poorly for IP analysis (data not shown). To improve specification for RPL30 IP, the mat3-4 transformants carrying the HA-tagged version of RPL30 (pRPL30-3XHA-Paro, mat3-4::RPL30-3XHA) were generated. Expression of the RPL30-3XHA protein in three independent mat3-4::RPL30-3XHA strains was confirmed by immunoblotting ( Figure 7C). RPL30-3XHA and its conjugates were first pulled down by anti-HA magnetic beads followed by IP using the anti-SUMO4 antibodyconjugated beads. Interestingly, a ladder-like migration pattern of RPL30-3XHA (presumably caused by the attachment of multiple SUMO4 molecules) was observed in cell lysates prepared from the mat3-4::RPL30-3XHA but not from the mat3-4 control cells ( Figure 7D). Intriguingly, the unmodified RPL30-3XHA protein was also pulled down by tandem IP.
To validate this result, a reciprocal purification was performed. SUMO4 conjugates were first immunopurified by the anti-SUMO4 antibody-conjugated beads and the SUMO4-conjugated RPL30 protein was further purified with anti-HA magnetic beads. Consistently, ladder-like slowly migrating RPL30-3XHA species were detected in the mat3-4::RPL30-3XHA but not in the mat3-4 control cells ( Figure 7E). Similarly, the unmodified RPL30-3XHA protein was the dominant protein purified by tandem IP. In both cases, four and five SUMO4 conjugates were found to be the dominant species of SUMOylated RPL30, even though no conspicuous poly-SUMOylated RPL30 was detected in the in vitro system ( Figure 7B). The discrepancy in the number of conjugated SUMO proteins between in vitro and in vivo experiments has also been reported previously by Duprez et al. (1999), Johnson and Blobel (1999), and Rodriguez et al. (2001). The absence of SUMO E3 ligase in the in vitro system may contribute to this discrepancy (Takahashi et al., 2003;Leung et al., 2015).
Even though SUMOylation sites were predicted in RPL30 (Supplemental Figure 8), mono-or di-SUMO conjugation cannot explain multiple slowly migrating species of RPL30-3XHA. Protein conjugation by poly-SUMO chains have been reported in budding yeast, mammals, and Trypanosoma brucei and it often leads to modified proteins with a ladder-like migration pattern (Tatham et al., 2001;Bylebyl et al., 2003;Liang et al., 2016;Iribarren et al., 2018). Furthermore, a conserved N-terminal SUMOylation site important for SUMO polymerization is shared by the SUMO proteins that are capable of forming the polymeric chains (Tatham et al., 2001;Bylebyl et al., 2003;Cheng et al., 2006;Iribarren et al., 2018). Interestingly, Chlamydomonas SUMO proteins also contain this N-terminal SUMOylation site ( Figure 7F; Wang et al., 2008;Shin et al., 2010), suggesting CrSUMO4 is able to form a poly-SUMO chain. Therefore, we suspect that RPL30 is conjugated with poly-SUMO4 chains. Whether CrSUMO4 is capable of forming poly-SUMO chains and whether RPL30 is conjugated with poly-SUMO4 chains remain to be determined.
Overexpression of the RPL30-SUMO4 DGG in mat3-4 Recapitulates the Size-Suppressing Phenotype of smt7-1 mat3-4 Because RPL30 was shown to be a target of SMT7 in vitro and in vivo, we hypothesized that SMT7 modulates the SUMOylation status of RPL30 to regulate cell division in a MAT3-dependent fashion. In smt7-1 mat3-4 cells, it is possible that a defect in SMT7 causes failure to replenish unconjugated RPL30, which is ratelimiting for cell division. To test this possibility, RPL30 level was knocked down in mat3-4 cells using an artificial microRNA (amiRNA) approach. If a relatively high amount of unconjugated RPL30 is critical for increased cell division number (and, thus, smaller cell size) in the mat3-4 mutant, it is predicted that mat3-4::amiRPL30 cells would have larger daughters than those of the mat3-4 strain. Down-regulation of RPL30 in mat3-4 strains by three different amiRNA constructs was confirmed in multiple independent transgenic lines by immunoblot analysis (Supplemental Figure 12). However, reducing RPL30 protein in mat3-4 cells failed to phenocopy the size suppression seen in the smt7-1 mat3-4 strain ( Table 3), suggesting that the amounts of unconjugated RPL30 are not rate-limiting for cell division.

DISCUSSION
In previous work, we isolated smt7-1 as a partial suppressor of the small size of mat3 mutant (Fang and Umen, 2008). Here, we characterized the SMT7 molecularly and biochemically and demonstrated that SMT7 encodes a bona fide Chlamydomonas SUMO protease. Key properties of SMT7 indicate that the SUMO protease activity of SMT7 is required to maintain cellular SUMO conjugation dynamics and that a defect in SMT7 increased SUMO4-conjugated species (Figure 6). Similar to yeast ULP1, human SENP1 and SENP2, and Arabidopsis ESD4 (Hang and Dasso, 2002;Zhang et al., 2002;Li and Hochstrasser, 2003;Panse et al., 2003;Xu et al., 2007), SMT7 was shown to be predominantly localized in discrete regions of nuclear membrane, presumably where the NPCs are. NPC-associated SUMO proteases are found to regulate SUMOylation of a gene-bound transcription factor (Texari et al., 2013), to maintain nucleoporin homeostasis (Chow et al., 2014), and to regulate repair of double strand DNA breaks (Duheron et al., 2017). Knowing that substrate specificity of SUMO proteases is strongly influenced by their restricted cellular localization, we propose that SMT7 regulates SUMO-conjugated proteins that transport to or from the nucleus.
It is already known that SUMO proteases play important roles in cell division control (Li and Hochstrasser, 1999;Pelisch et al., 2014;Schimmel et al., 2014;Eifler and Vertegaal, 2015). However, the identities of SUMO proteasemodulated target proteins and their links to the cell cycle functions remain incomplete. Despite SMT7 mRNA being a very low abundance gene, our data (Supplemental Figure 5) and the data of Zones et al. (2015) suggest that SMT7 mRNA is regulated diurnally with its expression elevated before onset of the S/M phase and gradually declines as cells exit mitosis. SMT7 protein, on the other hand, displayed dual expression peaks that are associated with Commitment and S/M phase (Figure 3), implicating the Modal cell sizes of mat3-4, smt7-1 mat3-4, and three independent mat3-4::amiRPL30 strains. SD values were derived from three to five independent experiments.
involvement of SMT7 in both size checkpoint functions. The incomplete correlation between expression patterns of SMT7 mRNA and its encoded protein suggests that posttranscriptional regulation of SMT7 may be important for its exerted function. Further support of SMT7 in size checkpoint function is evidenced by the requirement of its SUMO protease activity in the supernumerous cell divisions of mat3-4 cells. Unlike yeast ulp1-or ulp2-deficient cells that are either inviable or have abnormal cell shape and poor growth rate (Li andHochstrasser, 1999, 2000), the smt7-1 cells grow at a similar rate to wild-type cells and have normal-size daughters, and the smt7-1 mutation only affects size in the small-size mat3-4 cell background (Fang and Umen, 2008). This suggests that SMT7-dependent cellular activity is not ratelimiting for mitotic cell division when MAT3 is intact. It is possible that protease activities performed by other SUMO proteases maintain cell division when SMT7 is depleted. This possibility is supported by the presence of multiple SUMO proteases found in the Chlamydomonas genome (Breker et al., 2018). It will be interesting to determine whether other Chlamydomonas SUMO proteases act similarly to SMT7 to regulate cell division.
In this study we found that SMT7-dependent deconjugation of SUMOylated (SUMO4 in particular) targets affects size checkpoint control when MAT3 is defective. We found that RPL30 is apparently the prime SMT7 target in causing this phenotype and SMT7 is predominantly localized in NPCs. It is therefore reasonable to hypothesize that SMT7 removes SUMO4 from SUMO4-conjugated RPL30 in NPCs. SUMOylation of ribosome (A) Immunoblot detection of RPL30-SUMO4 DGG -3XHA protein from total cell lysates prepared from mat3-4 and five independent mat3-4::RPL30-SU-MO4 DGG -3XHA strains using the anti-HA antibody. The relative ratios of RPL30-SUMO4 DGG -3XHA/RbcL protein are shown. The ratio of RPL30-SUMO4 DGG -3XHA_B1/RbcL was arbitrarily set to one. (B) Immunoblot detection of SUMO4 DGG -RPL30-3XHA protein from total cell lysates prepared from mat3-4 and six independent mat3-4::SUMO4 DGG -RPL30-3XHA strains using the anti-HA antibody.
Here, we report that RPL30 is SUMOylated and demonstrate that RPL30 is conjugated with multiple SUMO4 molecules in Chlamydomonas. Because the N-terminal end of SUMO4 contains the conserved SUMOylation site, which allows conjugation with the poly-SUMO chains, and SUMOylated RPL30 formed a ladder-like pattern (Figure 8), we speculate that SUMO4 is capable of forming polymeric chains in Chlamydomonas. Poly-SUMO chains are reported to regulate resection of homology-dependent double strand break repair during vegetative growth and meiosis in budding yeast (Chen et al., 2016;Horigome et al., 2016) and to modulate nuclear foci formation and chromatin organization in procyclic Trypanosoma brucei (Iribarren et al., 2018). In addition, SUMO chains regulate the formation of the synaptonemal complex and promyelocytic leukemia nuclear bodies in mammals (Tatham et al., 2008;Watts and Hoffmann, 2011;Leung et al., 2015). The RPL30 SUMOylation process identified in this work may constitute the first step toward assessing the role of the potential polymeric SUMO structures in Chlamydomonas.
During the course of purifying the SUMO4-conjugated RPL30 protein, the unmodified RPL30 was repeatedly coimmunoprecipitated with SUMO4 conjugates. It is possible that the SUMO interacting motifs (SIM; Supplemental Figure 13A) of RPL30 provide the docking sites for poly-SUMO4 chains and/or poly-SUMO4 conjugates. In fact, SIM has been shown to be important for binding to poly-SUMO chains (Tatham et al., 2008;Iribarren et al., 2018;Sriramachandran et al., 2019). The requirement of SIM in RPL30 for binding to the poly-SUMO4 chain remains to be tested.
Overexpression of the SUMO protease-resistant RPL30-SUMO4 DGG protein, which mimics SUMOylated RPL30, but not overexpression of the RPL30 or SUMO4 protein in mat3-4 cells recapitulated smt7-1 mat3-4 and led to size suppression of the mat3-4 strain. These data provide compelling evidence to suggest that increased levels of SUMO4-conjugated RPL30 lead to altered cell size regulation in mat3-4 cells and to increased cell size. Even though RPL30 was found to be conjugated by multiple SUMO4 in vivo, an increase in RPL30-SUMO4 DGG level was sufficient to affect size in the mat3-4 strain. It will be of future interest to investigate how SUMOylation dynamics of RPL30 are regulated to control size-mediated cell cycle control.
How does an increase in SUMOylated RPL30 level affect cell division? RPL30 is one of the components constituting the large (60S) ribosomal subunit. RPL30 is localized on the surface of the large (60S) ribosomal subunit (Slavov et al., 2015) and is part of the intersubunit bridge that connects the large and small ribosomal subunits (Halic et al., 2005). The evolutionarily conserved residues involved in intersubunit bridges and RNA binding of RPL30 have been identified (Chen et al., 2003) and are conserved in Chlamydomonas RPL30 (Supplemental Figure 13B). It is therefore reasonable to propose that the SUMOylation status of RPL30 protein is involved in modulating RPL30-related ribosome biogenesis or preribosomal particle assembly, translocation, or maturation. The other possibility is that accumulation of SUMOylated RPL30 has an extra-ribosomal function. Many ribosomal proteins in plant and animal systems have been reported to have specialized functions other than translation (Warner and McIntosh, 2009;Nagaraj et al., 2016;Han et al., 2017;Li et al., 2017;Zhang et al., 2017;Dionne et al., 2019). Moreover, SUMO conjugation of mouse RPL11 protein has been shown to regulate p53-mediated stress responses (El Motiam et al., 2019), further supporting the importance of SUMOylation on the extraribosomal functions of ribosomal proteins. It will be of great interest to determine how SUMOylated RPL30 regulates cell division. The underlying cell-size control mechanism governed by the SMT7-RPL30-MAT3 module remains to be investigated.
It is important to note that overexpression of RPL30-SU-MO4 DGG -3XHA did not completely restore the size checkpoint defects caused by mat3-4, suggesting that other pathways acting downstream of MAT3/Rb are required for cell division. The mat3-4::RPL30-SUMO4 DGG -3XHA strain may be a valuable tool for investigating size control mechanism. Together, our findings shed light on a novel SMT7-RPL30-MAT3 module for size control.

RNA Isolation, RACE-PCR, RT-PCR, Quantitative RT-PCR, and Droplet Digital RT-PCR
Total RNA was isolated as previously described by Fang et al. (2014). Five micrograms of RNA was used for cDNA synthesis. cDNAs were reverse transcribed in the presence of a mixture of oligo dT and random primers (9:1) at 50°C for 50 min using ThermoScript Reverse Transcription Kit (Invitrogen) following the manufacturer's instructions. Each 20 mL PCR reaction contained 1 mL of 1/20-fold diluted cDNA, 1X Phusion GC buffer, 0.2 mM deoxynucleotide triphosphate (dNTP), 1 mM primers, 3% (v/v) DMSO, 0.5 M betaine, and 0.4 units Phusion High-Fidelity DNA polymerase (New England Biolabs). The PCR program was performed as follows: 98°C for 1 min, 50 cycles of 98°C for 10 s, 62°or 68°C for 10 s, and 72°C for 20 s per kb of the genes.
To isolate 39 ends of SMT7 cDNA, RACE-PCR was performed using the SMARTer RACE cDNA Amplification Kit (Clontech) following the manufacturer's instructions. Four to six micrograms of RNA was used for cDNA synthesis. The first strand cDNA was synthesized following the manufacturer's protocol. Approximately 850 bp SMT7 39 cDNA was amplified by SMT7 gene-specific primer (SMT7_39RACE) and Nested Universal Primer A (NUP, provided by the SMARTer RACE cDNA Amplification Kit) by Phusion High-Fidelity polymerase (New England Biolabs) and validated by sequencing (Supplemental Figure 2; Supplemental Table 6).
Because it was extremely difficult to obtain 59 RACE cDNA, we designed SMT7-specific primers and used RT-PCR to amplify SMT7 cDNA to validate the SMT7 gene model. Approximately 2.7 kb SMT7 cDNA was amplified by SMT7c_f8 and SMT7c_rXhoI (Supplemental Table 6). A 394bp SMT7 cDNA fragment flanking the start codon was amplified by SMT7c_f9 and SMT7c_r10 (Supplemental Table 6). The 20 mL PCR reaction contained 1 mL cDNA, 1X Phusion GC buffer, 0.2 mM dNTP, 1 mM of each primer, 3% (v/v) DMSO, 0.5 M betaine, 7.5% (v/v) ethylene glycol (Zhang et al., 2009), and 0.4 units Phusion High-Fidelity DNA polymerase (New England Biolads). The PCR program was performed as follows: 98°C for 1 min, 50 cycles of 98°C for 10 s, 62°to 65°C for 10 s, and 72°C for 20 s per kb of DNA.
Ten microliters of RT-qPCR reaction contained 2.5 mL of 1/20 diluted cDNA, 0.2 mM of primers, and 5 mL of 2X KAPA SYBR FAST master mix (KAPA Biosystems). The primers used for RT-qPCR are listed in Supplemental Table 7. Real-time RT-PCR was performed by a Bio-Rad CFX96 (Bio-Rad). The following PCR program was used for amplification: 95°C for 2 min, 40 cycles of 95°C for 5 s, and 63°C for 25 s. G PROTEIN BETA SUBUNIT-LIKE PROTEIN (GBLP) was used as an internal control. PCR was performed in triplicate.
For droplet digital RT-PCR (dd RT-PCR), 20 uL of PCR reaction contained 1 mL of 1/20 diluted cDNA, 900 nM of SMT7_1 and SMT7_2, 250 nM of SMT7_P probe, and 10 mL of 23 ddPCR Supermix for Probes (Bio-Rad). The primers used for ddRT-PCR are listed in Supplemental Table 7. PCR was performed using the QX200 Droplet Digital PCR System (Bio-Rad). The following PCR program was used for amplification: 95°C for 10 min, 40 cycles of 94°C for 30 s, and 60°C for 1 min, followed by 98°C for 10 min to deactivate enzyme. PCR was performed in triplicate.

Cell-size Measurement and Culture Synchronization
For cell-size measurement, a dark-shifted experiment was performed as previously described by Fang et al. (2014). Cell size was measured using Coulter Counter (Beckman Coulter) as previously described by Fang et al. (2006).
For culture synchronization, Chlamydomonas cells were cultured in HSM in a 12-h:12-h light:dark cycle with equal fluences of blue (450 nm) and red (630 nm) light (150 mmol photons m -2 s 21 each). Commitment and mitotic assays were performed as previously described by Fang et al. (2006).
During the course of our study, we found that some of the previously annotated SUMOylation components (Wang et al., 2008) are incorrect, which is likely caused by sequencing gaps and assembly errors of the previous genome. The updated annotations of the SUMOylation components, and their previous annotations were listed in Supplemental Table 8 as a reference.

Phylogenetic Analysis
Phylogenetic analysis was conducted using MEGA7 program . MUSCLE alignment of protein sequences from the SUMO protease domain of multiple SUMO proteases (Supplmental Data Set) were used to generate the phylogenetic tree. The unrooted phylogenetic tree (Supplemental Figure 2B) was constructed using the Maximum Likelihood method with 1000 bootstrap replications. The evolutionary distances were computed by Poisson of amino acid substitution model and the rate variation among sites was determined by gamma distribution. The heuristic to improve the likelihood of a tree was obtained by the Nearest-Neighbor-Interchange method.

Reconstitution of Chlamydomonas SUMOylation System and In Vitro SUMOylation Assay
The full-length cDNAs for SAE1, SAE2, SCE1, mature SUMO4, PCNA, and RPL30 were amplified by RT-PCR using the specific primer pairs listed in Supplemental Table 9. The SAE2 cDNA (2272 bp) was digested with BamHI and HindIII and cloned into the expression vector pCDFDuet (Promega) to generate pCDFDuet-SAE2 plasmid. The SAE1 cDNA (1193 bp) was then digested by BglII and XhoI and cloned into pCDFDuet-SAE2 to generate pCDFDuet-SAE1-SAE2 ( Figure 5A).
For protein induction, 0.2 mM isopropyl b-D-1-thiogalactopyranoside was added to the culture and the culture was incubated at 24°C for an additional 2.5 h with vigorous shaking. Cells (with the OD 600 reaching up to ;1.0) were harvested from 1 mL culture by centrifugation. Cell pellets were resuspended in 200 mL 13 SDS sampling buffer (10% [v/v]

Immunofluorescence
A linker sequence 59-GGCGGCAGCGGCGGCAGCGGCGGCAGCGGC GGC (encoding a peptide sequence: GGSGGSGGSGG) was introduced into the coding region between SMT7 and 3XHA by DNA synthesis (GENEWIZ); therefore, the pSMT7.1-lin-3XHA plasmid was generated. The pSMT7.1-lin-3XHA plasmid was transformed into 21gr and smt7-1 mat3-4 strains to generate SMT7-lin-3XHA and smt7-1 mat3-4::SMT7-lin-3XHA strains, respectively. One milliliter of cells (;10 6 cells mL 21 ) growing in the TAP medium was allowed to attach to polylysine-coated cover slips for 15 min at RT. The attached cells were fixed by immersion into two changes of pre-cooled 220°C 200 mL of 100% (v/v) methanol for 10 min and rehydrated in 500 mL of 13 PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 ) at RT for 15 min. Cover slips were then blocked in 150 mL blocking solution (5% [w/v] BSA and 5% [v/v] Normal Goat Serum [Jackson ImmunoResearch Laboratories] in 13 PBS) at RT for 90 min. After blocking, cover slips were incubated with anti-HA (Roche 3F10) antibody diluted at a ratio of 1:50 in blocking solution at 4°C overnight. Cover slips were washed three times with 13 PBST (13 PBS supplemented with 0.1% [v/v] Tween 20), 10 min each in a coplin jar at RT. Goat anti-rat Alexa Fluor 488conjugated secondary antibody (Thermo Fisher Scientific) was then diluted in blocking solution at 1:500 and incubated with cover slips for 120 min at RT in darkness. Following washing in 13 PBST for 10 min, cover slips were stained with DAPI (2 mg/mL in 13 PBS) for 10 min at RT. Following four washes with 13 PBST with 10 min per wash at RT, cover slips were mounted with Vectashield antifade medium (Vector laboratories H-1000). Zeiss 710 confocal microscope with filter sets (488 nm laser for Alexa Fluor 488; 405 nm UV light laser for DAPI) were used to capture the images.

Immunoprecipitation and Mass Spectrometry
The smt7-1 mat3-4 and smt7-1 mat3-4::SMT7 C928A -3XHA strains growing in HSM medium under continuous light were used for IP. Approximately 3 3 10 8 cells were harvested by centrifugation at 3220 3 g for 5 min at RT. Cells were washed two times with 10 mL of 13 PBS at RT. Cell pellets were then resuspended in 1 mL of 1X PBS containing 2 mM of disuccinimidyl glutarate (Thermo Fisher Scientific) or ethylene glycol bis(succinimidy succinate) (ThermoFisher Scientific) for cross-linking. Cells were incubated on ice for 30 min, and then quenched by adding Tris-HCl (pH 7.5) to 200 mM and incubated on ice for another 5 min. Cross-linked cells were washed twice with 10 mL 13 PBS at 4°C. Cell pellets were resuspended in 1 mL of RIPA buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 1% [v/v] Nonidet P-40, 5 mM EDTA, 5 mM EGTA, 10% [v/v] glycerol) containing 13 protease inhibitor cocktail (Sigma-Aldrich), 1 mM PMSF, 1 mM benzamidine, and 10 mM MG132 and vortexed for 45 min at 4°C in the presence of 500 mL zirconium beads (1 mm in diameter; Biospec). Cell lysates were transferred to a new 1.5-mL tube and centrifuged at 16,200 3 g for 30 min at 4°C to remove cell debris. Supernatant was then transferred to a new tube and protein concentration was determined using a Bio-Rad DC Protein Assay Kit.
Cell lysates were subjected to IP with anti-HA conjugated magnetic beads (Pierce). Thirty-five microliters (10 mg/ mL) of anti-HA magnetic beads (Pierce) were washed once with 1 mL 13 TBST and incubated with 3500 to 4000 mg protein lysates at 4°C for 4 h with 15 rpm end-over-end mixing. After incubation, magnetic beads were washed five times with 1 mL of RIPA buffer containing 13 protease inhibitor cocktail (Sigma-Aldrich), 1 mM PMSF, 1 mM benzamidine, and 10 mM MG132 for 10 min at 4°C. Immunoprecipitates were sent to the Proteomics Core Laboratory (Academia Sinica) for liquid chromatography-mass spectrometry analysis following the standard protocol (http://ipmb.sinica.edu.tw/proteomics/ Services.html; Shevchenko et al., 1996). Peptides and proteins were identified using Proteome Discoverer software (v.2.1, Thermo Fisher Scientific) with the SEQUEST search engine against Creinhardtii 281 v.5.5. protein database in Phytozome v.12.
For tandem immunopurification, 10 mg total cellular proteins were used for IP. IP was performed as described above except using the modified 13 RIPA buffer supplemented with 0.1% (w/v) SDS, 0.5% (w/v) sodium deoxycholate, 20 mM N-ethylmaleimide (NEM, SUMO protease inhibitor), 13 protease inhibitor cocktail (Sigma-Aldrich), 1 mM PMSF, 1 mM benzamidine, and 10 mM MG132. The immunoprecipitates from the first purification were eluted from magnetic beads using 2.5% (w/v) SDS solution. The eluate was diluted by adding 10 mL of modified RIPA buffer and concentrated using an Amicon Ultra-4 centrifugal filter unit (Merck) before secondary IP. Immunoprecipitates were then eluted by 2.5% (w/v) SDS solution and subjected to immunoblotting analysis. Horseradish peroxidaseconjugated anti-HA antibody (1:1000, Cell signaling, no. 2999) was used for immunoblotting.
Co-immunoprecipitation of SMT7 C928A -3XHA and RPL30 The cell lysates of smt7-1 mat3-4::SMT7 C928A -3XHA transformant expressing SMT7 C928A -3XHA protein were subjected to IP with anti-RPL30 antibody conjugated beads. The cell lysates of smt7-1 mat3-4 were used as a negative control. Approximately 40 mg anti-RPL30 antibody was conjugated to 30 mL of protein A Magnetic Sepharose beads (GE Healthcare). Briefly, RPL30 antibody was incubated with beads in 13 TBS binding buffer (20 mM Tris and 150 mM NaCl [pH 7.5]) at RT for 60 min. After 13 TBS binding buffer was removed, RPL30 antibody was cross-linked to beads by incubating in 500 mL of 200 mM triethanolamine (pH 8.9) supplemented with 50 mM dimethyl pimelimidate dihydrochloride (Sigma-Aldrich) at RT for 60 min. After the crosslinking solution was removed, 500 mL of 100 mM ethanolamine (pH 8.9) was added and incubated at RT for another 15 min to stop the crosslinking reaction. The un-bound antibody was removed by 500 mL of elution buffer (0.1 M Gly-HCl with 2 M Urea, pH2.9). The RPL30 antibody-conjugated beads were washed twice with 500 mL 13 TBS binding buffer and subjected to the IP procedure described previously with slight modifications. In brief, 1200 mg Chlamydomonas cell lysates were precleaned by incubating with 30 mL of protein A Magnetic Sepharose beads (GE Healthcare) at 4°C for 2 h. The precleaned cell lysates were then incubated with RPL30 antibody-conjugated beads supplemented with 0.1% (w/v) SDS and 0.5% (w/v) Sodium deoxycholate (SDC) at 4°C for 4 h.

Nuclei Isolation from Chlamydomonas
Nuclei were isolated as described previously (Ning et al., 2013) with slight modifications. Briefly, 2 3 10 8 cells were harvested from cultures growing in HSM medium under continuous illumination (GE F40 CW and GE F32T8) of 250 to 300 mmol photons m -2 s 21 at 24°C. The cell wall was removed by incubating cells in 20 mL of autolysin at RT overnight. Nonidet P-40 was used to test the effectiveness of autolysin treatment. Autolysin-treated cells were then harvested, resuspended, and incubated in 5 mL of 1X NIBA buffer (1X nuclei isolation buffer in CelLytic Plant Nuclei Isolation/ Extraction kit, Sigma-Aldrich, containing 1X protease inhibitor cocktail (Sigma-Aldrich), 1 mM PMSF, 1 mM benzamidine, 10 mM MG132, and 20 mM N-Ethylmaleimide (NEM)) on ice for 10 min. Homogenized cells were harvested by centrifugation at 1260 g for 10 min and resuspended in 10 mL 1X NIBA with 2% (v/v) Nonidet P-40 on ice for 60 min. Nuclei were harvested by centrifugation at 1000g for 30 min at 4°C. The harvested nuclear pellets were washed twice with 10 mL of 1X NIBA solution and resuspended in 1 mL RIPA buffer containing 1X protease inhibitor cocktail (Sigma-Aldrich), 1 mM PMSF, 1 mM benzamidine, 10 mM MG132, and 20 mM NEM.

Supplemental Data
Supplemental Figure 1. Nomarski images of daughter cells of the indicated strains.
Supplemental Figure 5. Expression pattern of SMT7 mRNA in a synchronized wild-type (21gr) culture (Fang et al., 2014) by droplet digital PCR.
Supplemental Figure 6. Clustal Omega alignment of protein sequences of Chlamydomonas SUMOylation components. Figure 7. Transcript abundance and protein alignment of Chlamydomonas SUMOs.
Supplemental Figure 10. Detection and identification of the immunoprecipitated RPL30 protein.
Supplemental Figure 13. Potential important features of RPL30 protein.
Supplemental Table 1. Modal daughter cell sizes (the size that appears most frequently in the cell size distribution data) of the indicated strains derived from crosses of smt7-1 mat3-4 pSMT7.1-3XHA #3 and 6145C (mt -).
Supplemental Table 2. Modal daughter cell sizes (the size that appears most frequently in the cell size distribution data) of the indicated strains.
Supplemental Table 3. Proteins identified by immunoprecipitation with the RPL30 antibody followed by Mass Spectrometry analysis.
Supplemental Table 5. The DNA sequences used for generating RPL30 artificial miRNA constructs.
Supplemental Table 6. RACE and RT-PCR primers used to amplify the SMT7 cDNA fragments. Table 7. Primers used for RT-PCR of the CDKB1 and RPL30, and for droplet RT-PCR of SMT7.

Supplemental
Supplemental Table 8. Nomenclature conversion table for annotation of SUMOylation components.
Supplemental Table 9. Primers used to amplify Chlamydomonas SUMOylation components for in vitro SUMOylation assay.
Supplemental Data Set. Alignments used to generate the phylogeny presented in Supplemental Figure 2B.